Shifting of the Overton window is the closest concept I know of, but it still seems too specific to fit this pattern. https://en.wikipedia.org/wiki/Overton_window

I'm always amazed at how much good science happens based on inferring things from what's known or guessed with high confidence.

The most powerful phrase in statistics is "suppose that it is..."

Why is that amazing? How would you expect it to work?

Did they? As far as I can recall 538 has always been clear that they deal in probabilities, not binary predictions. They gave him a small chance of winning for a reason, and it turned out chaos favored one of the low odds scenarios.

It's a gamble: stay with Slack and potentially be forced into an "Our Incredible Journey" migration in the future, or migrate now and shamble along until 40+ year old and profitable enterprise juggernaut Microsoft figures out how to make their product work right. Neither option is great.

Microsoft will never make Teams great, it will be just good enough that corporate bean counters won't let you pay for a competitor when you already get Teams "free".